Posted on 10/31/2014 3:20:01 PM PDT by scouter
I thought the end of October would be a good time to publish some new projections.
Here's an overview of the method I used.
So here are the new projections:
Scouter Ebola Projection Model Version 2.4 - Ebola Case Projections
*********************************************************
Projection Parameters
*********************************************************
Country: All Countries
Run Date/Time: 10/31/2014 at 17:49:43
Model: DTR Model
Start Date: 10/10/2014
End Date: 10/31/2014
Assumption of Reporting Completeness: Reported cases represent 100% of the true epidemic size
Rate of Increase per Day: 2.8%
*********************************************************
Starting and Ending Totals for Selected Time Period
Date Cases Deaths Daily New Cases Daily New Deaths
========== ==================== ==================== ==================== ====================
10/10/2014 8,692 4,254 148 113
10/31/2014 15,526 5,399 388 131
DTR 2.80% 1.14% 4.71% 0.72%
Weekly Projections for the Next 8 Weeks
Date Cases Deaths Daily New Cases Daily New Deaths
========== ==================== ==================== ==================== ====================
Weekly Projections for the Next 8 Weeks
11/07/2014 18,942 6,573 520 175
11/14/2014 23,025 7,942 634 212
11/21/2014 28,013 9,608 775 258
11/28/2014 34,110 11,636 948 315
12/05/2014 41,565 14,107 1,159 384
12/12/2014 50,684 17,118 1,418 468
12/19/2014 61,844 20,790 1,736 570
End of Month Projections for the Next Year from the End Date
Date Cases Deaths Daily New Cases Daily New Deaths
========== ==================== ==================== ==================== ====================
10/31/2014 15,597 7,972 2,073 514
11/30/2014 36,089 18,446 1,004 513
12/31/2014 87,106 44,524 2,456 1,256
01/31/2015 212,446 108,593 6,050 3,092
02/28/2015 478,450 244,563 13,715 7,010
03/31/2015 1,181,867 604,120 34,064 17,412
04/30/2015 2,847,331 1,455,437 82,395 42,117
05/31/2015 7,087,194 3,622,681 205,750 105,171
06/30/2015 17,172,560 8,777,904 499,783 255,468
07/31/2015 42,946,423 21,952,441 1,252,528 640,240
08/31/2015 107,596,017 54,998,649 3,143,581 1,606,869
09/30/2015 262,078,680 133,963,821 7,668,392 3,919,766
10/31/2015 658,460,530 336,577,891 19,292,735 9,861,651
bttt
Scouter Ebola ping.
If you would like to be added to or removed from my Ebola ping list, just send me a private FReep mail.
Interesting. Thanks for posting, running the numbers, and explaining your methodology. Health/life BUMP!
Oh, you know THAT's not gonna happen.
10/31/2015
658,460,530
336,577,891
19,292,735
9,861,651
For humanity's sake, I hope your model is as reliable as a Global Warming (TM) model.
PING!
Given the number of cases outside West Africa among the first 10,000 diagnosed Ebola cases, the potential for spread when there are 200k or more West African Ebola cases by January justifies a high level of concern.
Bttt.
Somehow 9 million deaths per day does not give me
a warm fuzzy feeling, or maybe it will as long as
I’m not bleeding from every orifice.
You need an earlier governor function to account for the fact that ebola victims are being shunned in many places where they would have been helped and thus spread the disease. The fact is no country or even no region in a country will achieve the percentage you are talking about, It did and probably does happen in remote villages where everyone touches the victims.
What you also need to do is account for the fact that Ebola is now endemic in some areas. Instead of burning itself out like it did in the past, it will simmer at a very low rate of victims and then pop out into 50-100 victims while the authorities scramble to contain that mini outbreak. Ebola used to follow the burnout curve but now it is too widespread to do that. However it will not follow your projections either because it dead ends in various locations and stops (while continuing in others).
An example for you is Nigeria. They had no idea that Sawyer had ebola when he was dying on the flight in, puking blood on the passenger next to him. She died, several flight attendants died and HCWs died. He infected 8 and killed 20, much higher than your linear / geometric spread assumption. However there were zero secondary infections, obviously much lower than your spread assumptions. You will need a much more complex model to account for that.
Infected 20 and killed 8.
The fact is no country or even no region in a country will achieve the percentage you are talking about, It did and probably does happen in remote villages where everyone touches the victims.
It happened with the Black Death in the Dark Ages.
Are you saying that is impossible?
It is possible in which case his estimates are LOW.
Lurking’
That was a great scene.
This stuff is really scary and our Governments don’t have a clue.
They reversed course and opened the slum. The slum residents are still mostly on their own. Here's a description: http://online.wsj.com/articles/liberian-slum-takes-ebola-treatment-into-its-own-hands-1414080932
Two months after Liberia’s largest slum fought a government-imposed Ebola quarantine, residents are in a desperate push to conquer the deadly virus—with or without the government’s help.
This is not Black Death that spread easily via nonhuman vectors. Mosquitoes probably don't carry it, nor ticks, fleas, etc. The proof of that is pretty simple, lots of people alive and well in the slum. To stay uninfected they simply need to avoid touching dead or dying victims. There is also probably a small amount of spread by infected victims before they get to the puking blood stage. But the residents have also stopped touching healthy-looking people as well.
However the endemic disease is a big concern. A human reservoir means it is ready to break out anywhere anytime as long as victim can get somewhere and humans can obviously do that. They will also send it here as long as we keep allowing them in. It needs to be contained there by finding and isolating the victims.
I think the effect of that is accounted for by using the transmission rate of the last 21 days. It is transmitting at its current rate, given all sorts of effects that I have no way of modeling, such as the weather, social interactions, whatever quarantines have been put into effect, etc.
The fact is no country or even no region in a country will achieve the percentage you are talking about...
I'm not making any predictions. The numbers I presented simply show the exponential progression at various points in time, based on the transmission rate of the last 21 days. Re-read point # 5 in my explanation (you did read it, didn't you?). I'm not saying everyone in Sierra Leone is going to die. But the current pool of infected patients in Sierra Leone will infect some number of other people. At some point someone from another country will be infected by a Sierra Leonean and will start a new outbreak in their country. But the cases in the new country will have their origins in Sierra Leone, in the same way that the outbreak in Dallas has its origins in Liberia. And the new outbreak will have an entirely new pool of potential victims.
What you also need to do is account for the fact that Ebola is now endemic in some areas. Instead of burning itself out like it did in the past, it will simmer at a very low rate of victims and then pop out into 50-100 victims while the authorities scramble to contain that mini outbreak. Ebola used to follow the burnout curve but now it is too widespread to do that. However it will not follow your projections either because it dead ends in various locations and stops (while continuing in others).
Good points. But I have neither the data nor the expertise to account for that explicitly. However, I think both points are at least partly accounted for by the two factors I mentioned above.
[Sawyer] infected 8 and killed 20, much higher than your linear / geometric spread assumption. However there were zero secondary infections, obviously much lower than your spread assumptions. You will need a much more complex model to account for that.
Not every case of Ebola spreads to others at the same rate. This model is not intended to be able to predict what will happen with individual cases. The fact that the cases "descended" from Sawyer do not match the projections in this model is not relevant. The model looks at what is happening at the macro level, not the micro level. It can't be denied that the epidemic has spread at something around 2.8% per day over the last 21 days. The model simply says that if nothing changes and it continues to spread at 2.8% per day, this is where it will lead. We know for a fact, though, that things will change. Some things will work to exacerbate the situation, and other things will work to ameliorate it. How they will balance out remains to be seen.
Despite its obvious deficiencies, I think this simple model helps to understand why Ebola is such a big deal. That's its intent. It is not intended to predict what's actually going to happen.
You seem relying on macro statistics (country numbers) to account for the micro fluctuations within each country. That's certainly valid in the short run as the characteristics of the pool of potential victims doesn't change that much. The complication comes when the victim pool changes. What I would expect is the small village victims (39 out of 40 dead in one village in a recent news story) morph into the Patrick Sawyer Lagos scenario (8 dead out of 5 million). The Nigerians were diligent (and authoritarian) and somewhat lucky. A less diligent and less authoritarian regime with less luck might have a very different outcome.
Some things will work to exacerbate the situation, and other things will work to ameliorate it. How they will balance out remains to be seen.
Agree completely but in general the ameliorating factors like removing and isolating patients and removing bodies are relatively easy in the cities. That was not the case in the villages which led to the statistics we have seen so far. The slums are in between those two extremes but it appears that the slum in West Point is kind of managing their outbreak. It's not going away there but not expanding 2.8% a day either. At that rate, in West Point they should have 8-9 times as many cases as they did in mid August when they closed it off and then reopened it. But as far as I can see, it is still about the same sporadic numbers of cases.
No, I'm relying on macro country-by-country numbers to account for the macro country-by-country numbers. My thought is that each country, at this point in the epidemic is relatively independent. They're implementing different policies, have different resources, and have had different levels of success in stopping the spread, and therefore different transmission rates. They should be considered separately. Their individual transmission rates account for many differences between them. But at some point, the virus is likely to spread to other countries which will also have their own transmission rates for the same reasons. At this point, Ebola outside of Liberia, Sierra Leone, and Guinea is negligible. Hopefully it will remain so. But when a significant outbreak occurs in another country I'll start tracking it independently and making separate projections.
Making the projections based on the growth that has occurred in the most recent 21 day period implicitly accounts, at least to a pretty significant degree, for the changes that occur in the pattern of spread over time. I have neither the data nor the resources to do better than that.
The Nigerians were diligent (and authoritarian) and somewhat lucky. A less diligent and less authoritarian regime with less luck might have a very different outcome.
That's part of the reason I think it's important to track the countries separately.
Agree completely but in general the ameliorating factors like removing and isolating patients and removing bodies are relatively easy in the cities.
I'm not so sure about that. There have been reports for weeks of bodies lying in the streets for days in cities in Liberia. It may be easier to take care of the bodies in Dallas or NYC, but perhaps not in Bombay or New Delhi or Hong Kong. Or in a war torn country. And even in Western countries, all it would take is a few dozen cases to overwhelm the isolation facilities we have, and then I suspect the patterns of spread would begin to take on some of the characteristics we see in countries with less sophisticated health care.
Don't forget that in our so-called superior health care system, two health care workers have already been infected. The fact that they survived can be chalked up to chance just as easily as it can be chalked up to the superior health care they received. I guarantee you that no hospital would not be able to sustain the level of care they received if that hospital had just 3 Ebola patients.
I've refined my projections over time, but even the early ones were reasonably in line with what actually occurred. If your goal is simply to show why Ebola is such a big deal to us even if it's still effectively contained in West Africa (and that's all my goal is), then I don't think there's all that much to be gained by taking into account all sorts of sophisticated concepts such as governor functions, weather, ability of western health care systems to absorb how many Ebola patients, etc. The point is that if we don't take Ebola seriously, it will get real bad real fast. That's what the average guy needs to understand hold the authorities' toes to the fire. If my projections are off by a factor of two, it really doesn't matter for that purpose.
We would probably become more authoritarian than Nigeria. The isolation facilities are comfy but tents are easy to put up and will isolate just as well. Our health care system would be overwhelmed with false positives which will cost billions in hospital protective measures, flying blood samples to CDC, state and local cleaning services, etc. At the same time care will deteriorate for other ailments. But we will not have victims at home vomiting blood on their relatives, home burials, etc. We will not have paranoia with people refusing to give up their sick relatives. Some infected may run themselves but they won't get far.
In short, it is not our first world health care that will save us, that will devolve quickly and we won't do much better than MSF. But our culture will quite obviously save us from pandemic.
Don't forget that in our so-called superior health care system, two health care workers have already been infected
2 out of 70 with protective measures that were worse than MSF. That compares pretty well to around 5-10% in third world. But it will get worse as you imply. The reason for that is that the false positives such as flu patients will overwhelm the system. Then the true positive ebola cases will not be handled properly until their tests come back in a day. That will cause infections in health care workers and other non-ebola patients.
All of this is easily prevented of course by stopping the 150 ebola land arrivals per day. If we have a handful of ebola land arrivals and American health care workers returning it will be fairly easy to track them aside from an occasional one who is obstinate like the Maine nurse but also infected (she is obviously not infected).
Liberia - Significant decline in Ebola cases. WP article dated November 3, 2014.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.